Product Recommended Using System Item-Based Collaborative Filtering With Slope One Algorithm Case Study: Omahgeulis.com
Authors: Deni Aditiya, Nur Nawaningtyas Pusparini, Rudi Setiyanto, Ahmad Syauki, Sofian Lusa
Number of views: 69
Sales of products online or e-commerce is currently becoming a trend in Indonesia. This is supported with the increasing number of internet users in Indonesia, making business opportunities in e-commerce more attractive to business people. PT.Victoria Care Indonesia is a company engaged in manufacturing for cosmetic and toiletris products, some brands of products are Victoria, Herborist, Miranda, Iria, Nuface, and Sixsence. In 2017 the company released Omahgeulis.com website to market its products online. But the number of products offered to make customers difficult to find products relevant to his interests so that sales at Omahgeulis.com less than the maximum. For that required a system that can facilitate customers in finding product information, one of them by recommending the product. Item-based collaborative filtering is a method for generating recommendations based on similarities between items. Slope One is an algorithm for rating predictions of products that have not been rated by users, this algorithm has advantages that are easy to implement, fast query time, and able to compete in accuracy with other approaches. The working principle of the Slope One algorithm is to calculate the average value of the deviation between items, the advantage is that when there is a new rating the system does not need to calculate from the beginning, but simply adds a new rating with the average deviation value then divides it by the amount of data so as to produce an average value for new deviation. To test the accuracy of the rating prediction value by the research system using Root Mean Squared Error (RMSE) by comparing the rating prediction value with the original rating value. From the test results obtained RMSE error value 0.87.